Table 3.
Simulation Study 1, Standardized Bias from Proration – Varied Item Means, Uniform Inter-Item Correlations
| Parameter | Items Per Scale |
Sample Size | MCAR Mechanism |
MAR Mechanism Due to External Variable |
MAR Mechanism Due to Complete Items on the Scale |
|---|---|---|---|---|---|
| Mean of X | 8 | 200 | 0.7012* | 0.7728* | 1.3568* |
| 500 | 1.1240* | 1.2362* | 2.1654* | ||
| 16 | 200 | 0.6906* | 0.7664* | 1.2285* | |
| 500 | 1.1466* | 1.2802* | 2.0496* | ||
| Mean of Y | 8 | 200 | 0.7387* | 0.8150* | 1.4230* |
| 500 | 1.1684* | 1.2916* | 2.2608* | ||
| 16 | 200 | 0.7166* | 0.7914* | 1.2754* | |
| 500 | 1.1125* | 1.2375* | 1.9812* | ||
| Variance of X | 8 | 200 | 0.1587 | 0.6583* | 1.4043* |
| 500 | 0.2602 | 1.1069* | 2.3601* | ||
| 16 | 200 | 0.0703 | 0.6150* | 1.2319* | |
| 500 | 0.1132 | 1.0062* | 2.0031* | ||
| Variance of Y | 8 | 200 | 0.1740 | 0.7312* | 1.5637* |
| 500 | 0.2680 | 1.1369* | 2.4294* | ||
| 16 | 200 | 0.0574 | 0.5861* | 1.1870* | |
| 500 | 0.0976 | 0.9467* | 1.9038* | ||
| Covariance | 8 | 200 | −0.0186 | 0.6930* | 0.3607 |
| 500 | −0.0124 | 1.1259* | 0.5993* | ||
| 16 | 200 | −0.0095 | 0.7182* | 0.3650 | |
| 500 | −0.0228 | 1.1464* | 0.5798* | ||
| Correlation | 8 | 200 | −0.0817 | 0.4814* | −0.1322 |
| 500 | −0.1103 | 0.7818* | −0.1942 | ||
| 16 | 200 | −0.0366 | 0.5282* | −0.0687 | |
| 500 | −0.0659 | 0.8415* | −0.1098 | ||
| Regression Coefficient |
8 | 200 | −0.0790 | 0.4670* | −0.1279 |
| 500 | −0.1072 | 0.7622* | −0.1841 | ||
| 16 | 200 | −0.0364 | 0.5022* | −0.0684 | |
| 500 | −0.0634 | 0.8075* | −0.1080 | ||
Note. The table contains standardized bias values for conditions with a 25% item-level missing data rate. The inter-item correlations were uniform but the item means varied.
Standardized bias values greater than or equal to .40 in absolute value are denoted by an asterisk (*).